regex-vs-llm-structured-text

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external text that could contain malicious instructions designed to manipulate LLM behavior during the validation step.
  • Ingestion points: The content parameter in the process_document and parse_structured_text functions in SKILL.md accepts raw, untrusted text.
  • Boundary markers: The prompt used in validate_with_llm interpolates data directly into the system message (Text: {original_text}) without using delimiters or instructions to ignore embedded commands.
  • Capability inventory: The skill performs data extraction and returns structured objects; it has no detected capabilities for file system access, shell command execution, or network exfiltration.
  • Sanitization: No validation or escaping of external content is implemented before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 09:24 AM